fix verify data(mulitk0)

example/01_gemm/gemm_xdl_skip_b_lds_fp16.cpp

@@ -87,10 +87,10 @@ template <typename DataType>
 std::ostream& show_2d_matrix(std::ostream& os, Tensor<DataType>& matrix)
 {
     os << "[" << std::endl;
-    for(int x = 0; x < matrix.mDesc.GetLengths()[0]; x++)
+    for(size_t x = 0; x < matrix.mDesc.GetLengths()[0]; x++)
     {
         os << "[";
-        for(int y = 0; y < matrix.mDesc.GetLengths()[1]; y++)
+        for(size_t y = 0; y < matrix.mDesc.GetLengths()[1]; y++)
         {
             os << std::setw(5) << static_cast<float>(matrix(x, y));
         }

include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_skip_b_lds_v1.hpp

@@ -191,7 +191,7 @@ struct GridwiseGemm_k0mk1_k0nk1_mn_xdlops_skip_b_lds_v1
 
         // 2-stage prefetch currently only support even number of K0 loop
         // TODO: add support for odd number of K0 loop
-        if(!((K0 / K0PerBlock) % 2 == 0))
+        if(!((K0 / K0PerBlock) % MultiK0 == 0))
         {
             return false;
         }